> -----Original Message-----
> From: Daniel P. Berrangé <berra...@redhat.com>
> Sent: Thursday, January 30, 2025 4:30 PM
> To: Shameerali Kolothum Thodi <shameerali.kolothum.th...@huawei.com>
> Cc: qemu-...@nongnu.org; qemu-devel@nongnu.org;
> eric.au...@redhat.com; peter.mayd...@linaro.org; j...@nvidia.com;
> nicol...@nvidia.com; ddut...@redhat.com; Linuxarm
> <linux...@huawei.com>; Wangzhou (B) <wangzh...@hisilicon.com>;
> jiangkunkun <jiangkun...@huawei.com>; Jonathan Cameron
> <jonathan.came...@huawei.com>; zhangfei....@linaro.org
> Subject: Re: [RFC PATCH 3/5] hw/arm/smmuv3: Associate a pci bus with a
> SMMUv3 Nested device
>
> On Fri, Nov 08, 2024 at 12:52:40PM +0000, Shameer Kolothum via wrote:
> > Subsequent patches will add IORT modifications to get this working.
> >
> > Signed-off-by: Shameer Kolothum
> <shameerali.kolothum.th...@huawei.com>
> > ---
> > hw/arm/smmuv3.c | 27 +++++++++++++++++++++++++++
> > include/hw/arm/smmuv3.h | 2 ++
> > 2 files changed, 29 insertions(+)
> >
> > diff --git a/hw/arm/smmuv3.c b/hw/arm/smmuv3.c
> > index 0033eb8125..9b0a776769 100644
> > --- a/hw/arm/smmuv3.c
> > +++ b/hw/arm/smmuv3.c
> > @@ -24,6 +24,7 @@
> > #include "hw/qdev-properties.h"
> > #include "hw/qdev-core.h"
> > #include "hw/pci/pci.h"
> > +#include "hw/pci/pci_bridge.h"
> > #include "cpu.h"
> > #include "trace.h"
> > #include "qemu/log.h"
> > @@ -2201,12 +2202,32 @@ static void smmu_realize(DeviceState *d,
> Error **errp)
> > smmu_init_irq(s, dev);
> > }
> >
> > +static int smmuv3_nested_pci_host_bridge(Object *obj, void *opaque)
> > +{
> > + DeviceState *d = opaque;
> > + SMMUv3NestedState *s_nested = ARM_SMMUV3_NESTED(d);
> > +
> > + if (object_dynamic_cast(obj, TYPE_PCI_HOST_BRIDGE)) {
> > + PCIBus *bus = PCI_HOST_BRIDGE(obj)->bus;
> > + if (s_nested->pci_bus && !strcmp(bus->qbus.name, s_nested-
> >pci_bus)) {
> > + object_property_set_link(OBJECT(d), "primary-bus", OBJECT(bus),
> > + &error_abort);
>
> Is the SMMUv3Nested useful if no 'primary-bus' is set ?
>
> If not, then the 'realize' method ought to validate 'pci-bus' is not
> NULL and and raise an error if NULL.
>
> After object_child_foreach_recursive returns, 'realize' should
> also validate that 'primary-bus' has been set, and raise an error
> it not set, to detect typos in the 'pci-bus' property.
I think that gets checked in the parent->realize()
smmu_base_realize()
if (s->primary_bus) {
pci_setup_iommu(s->primary_bus, &smmu_ops, s);
} else {
error_setg(errp, "SMMU is not attached to any PCI bus!");
}
I will double check if that covers all the corner cases or not.
Thanks,
Shameer
